Federal Reserve Bank of Dallas Globalization and Monetary Policy Institute Working Paper No. 296 https://doi.org/10.24149/gwp296 Capital Accumulation and Dynamic Gains from Trade * B. Ravikumar Ana Maria Santacreu Federal Reserve Bank of St. Louis Federal Reserve Bank of St. Louis Michael Sposi Federal Reserve Bank of Dallas January 2017 Abstract We compute welfare gains from trade in a dynamic, multi-country Ricardian model where international trade affects capital accumulation. We calibrate the model for 93 countries and examine transition paths between steady-states after a permanent, uniform trade liberalization across countries. Our model allows for both the relative price of investment and the investment rate to depend on the world distribution of trade barriers. Accounting for transitional dynamics, welfare gains are about 60 percent of those measured by comparing only the steady-states, and three times larger than those with no capital accumulation. We extend the model to incorporate adjustment costs to capital accumulation and endogenous trade imbalances. Relative to the model with balanced trade, the gains from trade increase more for small countries because they accumulate capital at faster rates by running trade deficits in the short run. JEL codes: E22, F11, O11 * Ana Maria Santacreu, Federal Reserve Bank of St. Louis, P.O. Box 442, St. Louis, MO 63166. 314-444- 6145. [email protected]. B. Ravikumar, Federal Reserve Bank of St. Louis, P.O. Box 442, St. Louis, MO 63166. 314-444-7312. b.r[email protected]. Michael Sposi, Federal Reserve Bank of Dallas, Research Department, 2200 N. Pearl Street, Dallas, TX 75201. 214-922-5881. This paper benefited from comments by Jonathan Eaton, Kim Ruhl, Mariano Somale, Felix Tintelnot, Kei-Mu Yi, and Jing Zhang. We are grateful to seminar audiences at Arizona State, Dallas Fed, Penn State, Purdue, UT Austin, and conference audiences at Empirical Investigations in International Trade, Midwest Macro, Midwest Trade, RIDGE Workshop on Trade and Firm Dynamics, UTDT Economics Conference, the Society for Economic Dynamics, and the System Committee for International Economic Analysis. The views in this paper are those of the authors and do not necessarily reflect the views of the Federal Reserve Bank of St. Louis, the Federal Reserve Bank of Dallas or the Federal Reserve System.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Federal Reserve Bank of Dallas Globalization and Monetary Policy Institute
Working Paper No. 296 https://doi.org/10.24149/gwp296
Capital Accumulation and Dynamic Gains from Trade*
B. Ravikumar Ana Maria Santacreu Federal Reserve Bank of St. Louis Federal Reserve Bank of St. Louis
Michael Sposi Federal Reserve Bank of Dallas
January 2017
Abstract We compute welfare gains from trade in a dynamic, multi-country Ricardian model where international trade affects capital accumulation. We calibrate the model for 93 countries and examine transition paths between steady-states after a permanent, uniform trade liberalization across countries. Our model allows for both the relative price of investment and the investment rate to depend on the world distribution of trade barriers. Accounting for transitional dynamics, welfare gains are about 60 percent of those measured by comparing only the steady-states, and three times larger than those with no capital accumulation. We extend the model to incorporate adjustment costs to capital accumulation and endogenous trade imbalances. Relative to the model with balanced trade, the gains from trade increase more for small countries because they accumulate capital at faster rates by running trade deficits in the short run.
JEL codes: E22, F11, O11
* Ana Maria Santacreu, Federal Reserve Bank of St. Louis, P.O. Box 442, St. Louis, MO 63166. 314-444-6145. [email protected]. B. Ravikumar, Federal Reserve Bank of St. Louis, P.O. Box 442, St. Louis, MO 63166. 314-444-7312. [email protected]. Michael Sposi, Federal Reserve Bank of Dallas, Research Department, 2200 N. Pearl Street, Dallas, TX 75201. 214-922-5881. This paper benefited from comments by Jonathan Eaton, Kim Ruhl, Mariano Somale, Felix Tintelnot, Kei-Mu Yi, and Jing Zhang. We are grateful to seminar audiences at Arizona State, Dallas Fed, Penn State, Purdue, UT Austin, and conference audiences at Empirical Investigations in International Trade, Midwest Macro, Midwest Trade, RIDGE Workshop on Trade and Firm Dynamics, UTDT Economics Conference, the Society for Economic Dynamics, and the System Committee for International Economic Analysis. The views in this paper are those of the authors and do not necessarily reflect the views of the Federal Reserve Bank of St. Louis, the Federal Reserve Bank of Dallas or the Federal Reserve System.
How large are the welfare gains from trade? This is an old and important question. This
question has been typically answered in static settings by computing the change in real
income from an observed equilibrium to a counterfactual equilibrium. In such computations,
the factors of production and technology in each country are held fixed and the change in real
income is entirely due to the change in each country’s trade share that responds to a change
in trade frictions. Recent examples include Arkolakis, Costinot, and Rodrıguez-Clare (2012)
who compute the welfare cost of autarky and Waugh and Ravikumar (2016) who compute
the welfare gains from frictionless trade.
By design, the above computations cannot distinguish between static and dynamic gains
from trade. We compute welfare gains from trade in a dynamic multi-country Ricardian
model where international trade affects the capital stock in each period. Our environment
is a version of Eaton and Kortum (2002) embedded into a two-sector neoclassical growth
model. There is a continuum of tradable intermediate goods. The technology for producing
the intermediate goods is country-specific and the productivity distribution is Frechet. Each
country is endowed with an initial stock of capital. Investment goods, produced using inter-
mediate goods, augment the stock of capital. Final consumption goods are also produced
using intermediate goods. Trade is subject to iceberg costs.
The model features two novel ingredients: (i) endogenous relative price of investment
and (ii) endogenous investment rate.
We compute the steady-state of the model for 93 countries and calibrate it to reproduce
the observed trade flows across countries, prices, and output per worker in each country in
2011. We use this steady-state as a benchmark and conduct a counterfactual exercise in
which trade barriers are reduced simultaneously in every country. We then compute the
transition path from the initial steady-state to the new steady-state. With this dynamic
path, we compute the welfare gains using a consumption equivalent measure as in Lucas
(1987).
We find that (a) comparing only steady states overstates the gains; the gains from trade
that include transition are about 60 percent of those measured by only comparing steady-
states, (b) static comparisons understate the gains; the dynamic gains with transition are
three times larger than than those measured in a static model with capital held fixed, and
(c) output increases on impact, but consumption drops since there is a large increase in
the marginal product of capital and a fall in the relative price of investment inducing large
2
increases in investment.
We then show the importance of the two main features of our model to analyze dynamic
welfare gains from trade. First, the endogenous relative price of investment allows countries
to attain permanently higher capital-output ratios, yielding higher output and consumption
in the steady-state. Second, the endogenous investment rate yields shorter half lives for
capital accumulation induced by temporarily high real rates of return to investment. As a
result, the model delivers large dynamic gains from trade.
The model’s predictions are consistent with several features of the data. Wacziarg and
Welch (2008) identify dates that correspond to a trade liberalization for 118 countries, and
show that, after such liberalization, GDP growth increases, the relative price of investment
falls fast and real investment rates increase. All these are features of our model. Further-
more, Wacziarg (2001) evaluates empirically several theories of dynamic gains from trade
in explaining the effect of trade on economic growth. Consistent with our results, he finds
that trade positively affects growth primarily through an increase in investment, and hence
capital accumulation.
Our solution method offers an efficient means to compute the transition path. The
method generalizes the algorithm of Alvarez and Lucas (2007b) to a dynamic environment
by iterating on a small subset of prices using information in excess demand equations. Such
an updating rule avoids computing costly gradients and typically converges in a matter of a
few hours on a basic laptop computer. This method applies to multi-country models of trade
with capital accumulation, CRRA preferences, linear depreciation of capital, and balanced
trade.
One limitation of our model, however, is that households cannot borrow against the
future, which restricts us to study transition paths in which investment is always positive.
To alleviate this limitation, we extend the baseline model by adding adjustment costs to
capital accumulation and endogenous trade imbalances. We provide a modified algorithm to
compute the transition paths in this economy as well.1
Using the extended model with adjustment costs, we consider the same reduction in
trade barriers as in the baseline model both when trade is balanced and when there are
endogenous trade imbalances. We find that countries that have a higher marginal product
of capital in the baseline model, grow faster, and in the extended model experience inflows
1Our algorithms for the baseline model and for the extended model rely on gradient-free updating rules.These methods are computationally less demanding than nonlinear solvers used by recent papers that studycapital accumulation and endogenous trade imbalances such as Eaton, Kortum, Neiman, and Romalis (2016)and Kehoe, Ruhl, and Steinberg (2016).
3
of capital and run a trade deficit early on. These countries then converge to a steady-state
with a trade surplus. On the contrary, slow-growing countries run a trade surplus early on,
but converge to a steady-state with a trade deficit. Welfare is slightly higher in the model
with trade imbalances than in the model with balanced trade for all countries. However,
relative to a model with balanced trade, countries that run a trade deficit early on exhibit
proportionately larger dynamic gains from trade than countries that run a surplus early on.
Our paper relates to several strands of literature. First, it relates to two recent studies
that examine dynamic trade models. Eaton, Kortum, Neiman, and Romalis (2016) and
Caliendo, Dvorkin, and Parro (2015) compute the transitional dynamics of an international
trade model by computing period-over-period change in endogenous variables as a result of
changes in parameters (this is the so-called hat-algebra approach). Our approach differs from
theirs in several aspects. First, we solve for the transition of our model in levels; we do not
use hat algebra. By solving the model in levels, we are able to validate the cross-sectional
predictions of our model. In particular, we find that our model is consistent with the cross-
sectional distribution of capital and investment rates in the data. Second, computing the
initial steady-state in levels allows us to impose discipline on the type of trade liberalization
we are interested in, which is not possible without knowing the initial levels. Finally, Eaton,
Kortum, Neiman, and Romalis (2016) solve for the planner’s problem and assume that the
Pareto weights remain constant across counterfactuals, implying that each country’s share
in world consumption is fixed across counterfactuals. In our computation, however, each
country’s share in world consumption changes across counterfactuals and along the transition
path, a feature that is important when studying welfare.2
A second strand of literature has incorporated capital accumulation into trade models to
study welfare gains from trade. Alvarez and Lucas (2007a) develop a method that approx-
imates the dynamics by linearizing around the steady-state. Alessandria, Choi, and Ruhl
(2015) consider welfare gains from trade in a two-country model with capital accumulation
and highlight short run frictions such as fixed and sunk costs to export. Finally, Brooks and
Pujolas (2016) compare dynamic welfare gains in a model with endogenous capital accumu-
lation to those in a static model. They do so in the context of a two-country model with
balanced trade.
Anderson, Larch, and Yotov (2015) study the transitional dynamics via capital cumula-
2Zylkin (2016) uses an approach similar to “hat algebra” to study how China’s integration from 1993-2011has had an effect on investment and capital accumulation in the rest of the world. His “hat algebra” approachdiffers from other papers in that he computes the change of the variable from its baseline equilibrium valueto its counterfactual equilibrium value, rather than computing period-over-period changes.
4
tion in a multi-country model to measure the welfare gains from trade. Our paper builds on
their work by incorporating into the model a relative price of investment and endogenous
investment rate that each depend on the world distribution of trade barriers. These features
imply that anticipated changes to future trade costs have an impact on current consump-
tion and saving decisions. The endogenous relative price implies that countries can attain
higher steady-state capital stocks. The endogenous investment rate implies that countries
accumulate capital more quickly in response to a trade liberalization. Both of these features
affect the computed gains from trade and are consistent with empirical evidence as shown
by Wacziarg (2001) and Wacziarg and Welch (2008).
Our extended model adds to a strand of literature that analyzes dynamics in interna-
tional trade via endogenous trade imbalances. Reyes-Heroles (2016) studies endogenous
trade imbalances in a multi-country model without capital, while Sposi (2012) studies en-
dogenous trade imbalances with an exogenous nominal investment rate for capital. Kehoe,
Ruhl, and Steinberg (2016) combine capital accumulation and endogenous trade imbalances
into a two-country, general equilibrium model of trade.
Finally, recent studies have used “sufficient statistics” approaches to measure changes in
welfare by looking at changes in the home trade share (Arkolakis, Costinot, and Rodrıguez-
Clare, 2012). In our baseline model the sufficient-statistics formula is only valid across
steady-states, but not along the transition path. We show that measuring changes in welfare
using changes in consumption along the transition path yields very different implications than
one would obtain by using sufficient statistics. Moreover, in our extended model we show that
the sufficient-statistics formula breaks down even across steady-states, with systematically
larger errors for countries that run steady-state trade deficits compared to countries that
run surpluses. That is, when trade imbalances are endogenously determined, changes in the
home trade share are not sufficient to characterize the changes in welfare, or in income for
that matter, across steady-states.
The rest of the paper proceeds as follows. Section 2 presents the model. Section 3
describes the quantitative exercise. Section 4 reports the counterfactuals, and section 5
concludes.
2 Model
There are I countries indexed by i = 1, . . . , I and time is discrete, running from t = 1, . . . ,∞.
There are three sectors: consumption, investment, and intermediates, denoted by c, x, and
5
m respectively. Neither consumption goods nor investment goods are tradable. There is a
continuum of intermediate varieties that are tradable. Production of all the goods are carried
out by perfectly competitive firms. As in Eaton and Kortum (2002), each country’s efficiency
in producing each intermediate variety is a realization of a random draw from a country-
and time-specific distribution. Trade in intermediate varieties is subject to iceberg costs.
Each country purchases each intermediate variety from its lowest-cost supplier and all of the
varieties are aggregated into a composite intermediate good. The composite intermediate
good, which is nontradable, is used as an input along with capital and labor to produce the
consumption good, the investment good, and the intermediate varieties.
Each country has a representative household. The representative household owns its
country’s stock of capital and labor, which it inelastically supplies to domestic firms, and
purchases consumption and investment goods from the domestic firms.
2.1 Endowments
In each period, the representative household in country i is endowed with a labor force of
size Li, which is constant over time, and a stock of capital in the initial period, Ki1.
2.2 Technology
There is a unit interval of varieties in the intermediates sector. Each variety within the
sector is tradable and is indexed by v ∈ [0, 1].
Composite good Within the intermediates sector, all of the varieties are combined
with constant elasticity in order to construct a sectoral composite good according to
Qit =
[∫ 1
0
qit(v)1−1/ηdv
]η/(η−1)
where η is the elasticity of substitution between any two varieties.3 The term qit(v) is the
quantity of good v used by country i to construct the composite good at time t and Qit is the
quantity of the composite good available in country i to be used as an intermediate input.
3The value η plays no quantitative role other than satisfying technical conditions which ensure convergenceof the integrals.
6
Varieties Each variety is produced using capital, labor, and the composite intermediate
good. The technologies for producing each variety are given by
Ymit(v) = zmi(v)(Kmit(v)αLmit(v)1−α)νmMmit(v)1−νm
The term Mmit(v) denotes the quantity of the composite good used by country i as an input
to produce Ymit(v) units of variety v, while Kmit(v) and Lmit(v) denote the quantities of
capital and labor employed.
The parameter νm ∈ [0, 1] denotes the share of value added in total output, while α
denotes capital’s share in value added. Each of these coefficients is constant both across
countries and over time.
The term zmi(v) denotes country i’s productivity for producing variety v. Following
Eaton and Kortum (2002), the productivity draw comes from independent country-specific
Frechet distributions with shape parameter θ and country-specific scale parameter Tmi, for
i = 1, 2, . . . , I. The c.d.f. for productivity draws in country i is Fmi(z) = exp(−Tmiz−θ).In country i the expected value of productivity across the continuum is γ−1T
1θmi, where γ =
Γ(1 + 1θ(1− η))
11−η and Γ(·) is the gamma function, and T
1θmi is the fundamental productivity
in country i.4 If Tmi > Tmj, then on average, country i is more efficient than country j at
producing intermediate varieties. The parameter θ > 0 governs the coefficient of variation
of the efficiency draws. A larger θ implies more variation in efficiency across countries and,
hence, more room for specialization within each sector; i.e., more trade.
Consumption good Each country produces a consumption good using capital, labor,
and intermediates according to
Ycit = Aci(KαcitL
1−αcit
)νcM1−νc
cit
The terms Kcit, Lcit, and Mcit denote the quantity of capital, labor, and composite interme-
diate good used by country i to produce Ycit units of consumption at time t. The parameters
α and νc are constant across countries and over time. The term Aci captures country i’s
productivity in the consumption goods sector—this term varies across countries.
4As discussed in Waugh (2010) and Finicelli, Pagano, and Sbracia (2012), fundamental productivity differsfrom measured productivity because of selection. In a closed economy, country i produces all varieties in thecontinuum so its measured productivity is equal to its fundamental productivity. In an open economy, countryi produces only the varieties in the continuum for which it has a comparative advantage and imports therest. So its measured productivity is higher than its fundamental productivity, conditioning on the varietiesthat it produces in equilibrium.
7
Investment good Each country produces an investment good using capital, labor, and
intermediates according to
Yxit = Axi(KαxitL
1−αxit
)νxM1−νx
xit
The terms Kxit, Lxit, and Mxit denote the quantity of capital, labor, and composite interme-
diate good used by country i to produce Yxi units of investment at time t. The parameters
α and νx are constant across countries and over time. The term Axi captures country i’s
productivity in the investment goods sector—this term varies across countries.
2.3 Trade
International trade is subject to barriers that take the iceberg form. Country i must purchase
dij ≥ 1 units of any intermediate variety from country j in order for one unit to arrive; dij−1
units melt away in transit. As a normalization we assume that dii = 1 for all i.
2.4 Preferences
The representative household values consumption per capita over time, Cit/Li, according to
∞∑t=1
βt−1Li(Cit/Li)
1−1/σ
1− 1/σ
where β ∈ (0, 1) denotes the period discount factor and σ denotes the intertemporal elasticity
of substitution. Both parameters are constant across countries and over time.
Capital accumulation Each period the representative household enters with Kit units
of capital, which depreciates at the rate δ. Investment, Xit, adds to future capital.
Kit+1 = (1− δ)Kit +Xit
Budget constraint The representative household earns income by supplying capital,
Kit, and labor, Li, inelastically to domestic firms earning a rental rate rit on each unit of
capital and a wage rate wit on each unit of labor. The household purchases consumption
at the price Pcit per unit and purchases investment at the price Pxit per unit. The period
8
budget constraint is given by
PcitCit + PxitXit = ritKit + witLi
2.5 Equilibrium
A competitive equilibrium satisfies the following conditions: i) taking prices as given, the
representative household in each country maximizes its lifetime utility subject to its budget
constraint and technology for accumulating capital, ii) taking prices as given, firms maximize
profits subject to the available technologies, iii) intermediate varieties are purchased from
their lowest-cost provider subject to the trade barriers, and iv) all markets clear. At each
point in time, we take world GDP as the numeraire:∑
i ritKit + witLi = 1 for all t. We
describe each equilibrium condition in more detail in Appendix A.
2.6 Welfare Analysis
We measure welfare using consumption-equivalent units since utility in our model is defined
over consumption. This is a departure from much of the literature on static models in which
welfare gains are computed as changes in income. In a dynamic model, as income changes
along the transition path we need to examine how the income is allocated to consumption
and investment.
We follow Lucas (1987) and compute the constant, λdyni :
∞∑t=1
βt−1Li
((1 +
λdyni
100
)C?i /Li
)1−1/σ
1− 1/σ=∞∑t=1
βt−1Li
(Cit/Li
)1−1/σ
1− 1/σ(1)
where C?i is the (constant) consumption in the benchmark steady-state in country i, and Cit
is the consumption in the counterfactual at time t.5 We refer to λdyni as “dynamic gains.”
In steady-state, this formula can be expressed as
⇒ 1 +λssi100
=C??i
C?i
(2)
where C??i is the consumption in the the new (counterfactual) steady-state in country i. In
5We calculate sums using the counterfactual transition path solved from t = 1, . . . , 150 and then set thecounterfactual consumption equal to the new steady-state level of consumption for t = 151, . . . , 1000.
9
this expression, λssi measures gains from trade across steady-states in country i. In our model
consumption is proportional to income across countries in the steady-state.6
Dynamic welfare gains require us to compute the entire transition path for consumption,
which depends on the transition for capital accumulation. The dynamics of capital are
governed by the Euler equation. In particular, combining the Euler equation with the budget
constraint and the capital accumulation technology, the equilibrium law of motion for capital
must obey the following equation in every country(1 +
rit+1
Pxit+1
− δ)(
Pxit+1
Pcit+1
)Kit+1 +
(wit+1
Pcit+1
)Li −
(Pxit+1
Pcit+1
)Kit+2
= βσ(
1 +rit+1
Pxit+1
− δ)σ (
Pxit+1/Pcit+1
Pxit/Pcit
)σ×[(
1 +ritPxit− δ)(
PxitPcit
)Kit +
(witPcit
)Li −
(PxitPcit
)Kit+1
]This is the key equation to analyze dynamics, and it constitutes the main departure from
the existing models analyzing welfare gains from trade.7 Note that the dynamics of capital in
country i depend on the capital stocks in all other countries since the prices are determined
in the world economy due to trade. Thus, the dynamics are pinned down by the solution
to a system of I second-order, nonlinear difference equations. The optimality conditions for
the firms combined with the relevant market clearing conditions pin down the prices as a
function of the capital stocks across countries.
3 Quantitative exercise
We describe in Appendix B the details of our algorithm for computing the dynamic equi-
librium in the baseline model. Broadly speaking, we first reduce the infinite dimension of
the problem down to a finite-time model with t = 1, . . . , T periods. We make T sufficiently
large to ensure convergence to a new steady-state. As such, this requires us to first solve
for a terminal steady-state to use as a boundary condition for the path of capital stocks. In
addition, we take initial capital stocks as given by the initial steady-state.
In steady-state, all endogenous variables are constant over time. Table B.1 provides the
6The formula for to ratio of consumption to income in country i is CiyiLi
= 1− αδ1β−(1−δ) .
7Anderson, Larch, and Yotov (2015) use a similar expression to measure dynamic welfare gains fromtrade, but they impose assumptions on preferences and technologies that yield a fixed investment rate. Asa result, their model does not admit an Euler equation.
10
equilibrium conditions that describe the solution to the steady-state in our model. Our tech-
nique for computing the steady-state equilibria are standard, while our method for computing
the equilibrium transition path between steady-states is new.8
3.1 Calibration
We calibrate the initial parameters of the model to match data in 2011. Our assumption is
that the world is in steady-state at this time. Our model covers 93 countries (containing 91
individual countries plus 2 regional country groups). Table F.1 in the Appendix provides a
list of the countries along with their 3-digit ISO codes. This set of countries accounts for 90
percent of world GDP as measured by the Penn World Tables, and for 84 percent of world
trade in manufactures as measured by the United Nations Comtrade Database. Appendix
E provides the details of our data.
Common parameters The values for the common parameters are reported in Table
1. We use recent estimates of the trade elasticity by Simonovska and Waugh (2014) and set
θ = 4. The value for η plays no quantitative role in the Eaton-Kortum model of trade other
than satisfying the condition that 1 + 1θ(1− η) > 0; we set η = 2.
Table 1: Common parameters
θ Trade elasticity 4η Elasticity of substitution between varieties 2α Capital’s share in value added 0.33β Annual discount factor 0.96δ Annual depreciation rate for stock of capital 0.06σ Intertemporal elasticity of substitution 0.67νc Share of value added in final goods output 0.91νx Share of value added in investment goods output 0.33νm Share of value added in intermediate goods output 0.28
8We solve for the competitive equilibrium of the model. This differs from Eaton, Kortum, Neiman,and Romalis (2016), who solve the planner’s problem. In particular, they use the social planner’s problemto solve for trade imbalances using fixed weights across counterfactuals. This implies that each countrysshare in world consumption expenditures (i.e., the numeraire in their setting) is fixed across counterfactuals.In a decentralized economy, these shares would change, and still be efficient. We see this in our owncounterfactuals. The second welfare theorem states that any social planner outcome can be replicated ina decentralized market with the appropriate transfers. In our context, this implies that the social plannerweights would need to change in order to generate the same allocation as the decentralized economy withouttransfers (i.e., in our counterfactuals).
11
In line with the literature, we set the share of capital in value added to α = 0.33 (from
Gollin, 2002), the discount factor β = 0.96, and the intertemporal elasticity of substitution
σ = 0.67.
We compute νm = 0.28 by taking the cross-country average of the ratio of value added to
gross output of manufactures. We compute νx = 0.33 by taking the cross-country average of
the ratio of value added to gross output of investment goods. Computing νc is slightly more
involved since there is not a clear industry classification for consumption goods. Instead, we
infer this share by interpreting national accounts data through the lens of our model. We
begin by noting that, from combining firm optimization and market clearing conditions for
capital and labor we get
riKi =α
1− αwiLi
In steady-state, the Euler equation and the capital accumulation technology imply
PxiXi =δα
1β− (1− δ)
wiLi1− α
= φxwiLi1− α
We compute φx by taking the cross-country average of the share of gross fixed capital for-
mation in nominal GDP. Given this value, and the relation φx = δα1β−(1−δ) , the depreciation
rate for capital is δ = 0.06. The household’s budget constraint then implies that
PciCi =wiLi1− α
− PxiXi = (1− φx)wiLi1− α
Consumption in our model corresponds to the sum of private and public consumption,
changes in inventories, and net exports. We can use the trade balance condition together
with the firm optimality and the market clearing conditions for sectoral output to obtain
PmiQi = [(1− νx)φx + (1− νc)(1− φx)]wiLi1− α
+ (1− νm)PmiQi (3)
where PmiQi is total absorption of manufactures in country i and wiLi1−α is the nominal GDP.
We use a standard method of moments estimator to back out νc from equation (3).
Country-specific parameters We set the workforce, Li, equal to the total popula-
tion. The remaining parameters Aci, Tmi, Axi and dij, for (i, j) = 1, . . . , I, are not directly
observable. We parsimoniously back these out by linking structural relationships of the
model to observables in the data.
12
The equilibrium structure relates the unobserved trade barrier for any given country pair
directly to the ratio of intermediate-goods prices in the two countries and the trade shares
between them as followsπijπjj
=
(PmjPmi
)−θd−θij (4)
Appendix E provides the details for how we construct the empirical counterparts to prices
and trade shares. For observations in which πij = 0, we set dij = 108. We also set dij = 1 if
the inferred value of trade cost is less than 1.
Lastly, we derive three structural relationships to pin down the productivity parameters
Aci, Tmi, and Axi.
Pci/PmiPcU/PmU
=
(Tmiπii
) 1θ/Aci(
TmUπUU
) 1θ/AcU
(Tmiπii
) 1θ
(TmUπUU
) 1θ
νc−νmνm
(5)
Pxi/PmiPxU/PmU
=
(Tmiπii
) 1θ/Axi(
TmUπUU
) 1θ/AxU
(Tmiπii
) 1θ
(TmUπUU
) 1θ
νx−νmνm
(6)
ymiymU
=
(AciAcU
)(AxiAxU
) α1−α
(Tmiπii
) 1θ
(TmUπUU
) 1θ
1−νc+ α
1−α (1−νx)νm
(7)
The three equations relate observables—the price of consumption relative to intermediates,
the price of investment relative to intermediates, income per capita, and home trade shares—
to the unknown productivity parameters. These derivations are in Appendix D. We set
AcU = TmU = AxU = 1 as a normalization, where the subscript U denotes the U.S. For each
country i, system (5)–(7) yields three nonlinear equations with three unknowns: Aci, Tmi,
and Axi. Information about constructing the empirical counterparts to Pci, Pmi, Pxi, πii and
ymi is available in Appendix E.
These equations are quite intuitive. The expression for income per capita provides a
measure of aggregate productivity across all sectors: higher income per capita is associated
with higher productivity levels, on average. The two expressions for relative prices tell us
how to allocate the productivity across sectors.
The expressions for relative prices boil down to two components. The first term reflects
something akin to the Balassa-Samuelson effect: All else equal, a higher price of capital
13
relative to intermediates suggests a low productivity in capital goods relative to intermediate
goods. In our setup, the productivity for the traded intermediate good is endogenous,
reflecting the degree of specialization as captured by the home trade share. The second term
reflects the extent to which the two goods utilize intermediates with different intensities.
If measured productivity is relatively high in intermediates, then the price of intermediate
input is relatively low and the sector that uses intermediates more intensively will, all else
equal, have a lower relative price.
3.2 Model fit
Our model consists of 8832 country-specific parameters: I(I − 1) = 8556 bilateral trade
barriers, (I − 1) = 92 consumption-good productivity terms, (I − 1) = 92 investment-good
productivity terms, and (I − 1) = 92 intermediate-goods productivity terms.
Calibration of the country-specific parameters utilizes 8924 independent data points. The
trade barriers use up I(I − 1) = 8556 data points for bilateral trade shares and (I − 1) = 92
for ratio of absolute prices of intermediates. The productivity parameters use up (I−1) = 92
data points for the price of consumption relative to intermediates, (I − 1) = 92 for the price
of investment relative to intermediates, and (I − 1) = 92 for income per capita.
As such, there 92 more data points than parameters so our model does not perfectly
replicate the data. Another way to interpret this is that there is one equilibrium condition
for each country that we did not impose on our identification:
Pmi = γ
[I∑j=1
(umjdij)−θTmj
]− 1θ
The model matches the targeted data well. The correlation between model and data
is 0.96 for the bilateral trade shares, 0.97 for the absolute price of intermediates, 1.00 for
income per capita, 0.96 for the price of consumption relative to intermediates, and 0.99 for
the price of investment relative to intermediates.
Indeed, since we utilized relative prices of consumption and investment, not the absolute
prices, matching the absolute prices is a test of the model. The correlation between model
and data is 0.93 for the absolute price of consumption, and 0.97 for the absolute price of
investment.
14
Figure 1: Model fit: The vertical axis represents the model and the horizontal axis representsthe data
(a) Capital-labor ratio
1/256 1/64 1/16 1/4 1 41/256
1/64
1/16
1/4
1
4
ARM
AUSAUT
BDI
BEN
BGD
BGRBHSBLR
BLZ
BRA
BRBBTN
CAF
CAN
CHE
CHL
CIV
CMR
COL
CPV
CRI
CYPCZE
DEUDNK
DOMECU
EGY
ESP
ETH
FIN
FJI
FRAGBR
GEO
GRC
GTMHND
HUN
IDNIND
IRL
IRN
ISLISRITA
JAM
JOR
JPN
KGZ
KHM
KOR
LKA
LSO
MAR
MDA
MDG
MDV
MEX
MKD
MOZ
MUS
MWI
NPL
NZL
PAK
PER
PHL
POL
PRT
PRY
ROURUS
RWASENSTP
SWE
THATUN
TUR
TZAUGA
UKR
URY
USA
VCT
VEN
VNMYEM
ZAF BALCHM
45o
(b) Investment rate
1/8 1/4 1/2 1 2 41/8
1/4
1/2
1
2
4
ARM
AUSAUT
BDI
BEN
BGDBGR
BHSBLR
BLZ
BRA
BRB
BTN
CAF
CANCHE
CHL
CIV
CMR
COL CPV
CRI
CYPCZEDEUDNK
DOM ECUEGY
ESP
ETH
FIN
FJI
FRAGBR
GEO
GRC
GTMHND
HUN
IDN
IND
IRL
IRNISLISRITA
JAM
JOR
JPN
KGZ
KHMKOR
LKA
LSO
MAR
MDA
MDG
MDV
MEX
MKDMOZ MUS
MWI
NPLNZLPAK PER
PHLPOL
PRT
PRY
ROU
RUSRWASEN
STP
SWETHA
TUNTUR
TZAUGA
UKR
URYUSA
VCT
VEN
VNM
YEM
ZAFBAL
CHM
45o
Implication for capital stock In our calibration we targeted income per capita. The
burden is on the theory to disentangle what fraction of the cross-country income gap can be
attributed to differences in capital and what fraction to differences in TFP.
Figure 1 shows that the model matches the data on capital-labor ratios across countries
quite closely: the correlation is 0.93. It also shows that our model captures well the invest-
ment rate, XiyiLi
, across countries in 2011. Note that we are imposing steady-state in 2011,
which implies that the investment rate is proportional to the capital-output ratio. Since our
model matches GDP by construction, and also does well explaining capital stocks, our ability
to replicate the investment rate is limited to the extent that the steady-state assumption is
valid in the data.
4 Counterfactuals
In this section we implement a counterfactual trade liberalization via a one time, permanent
reduction in trade barriers. The world begins in the calibrated steady-state. At the beginning
of period t = 1, trade barriers fall uniformly across all countries such that the ratio of world
trade to GDP increases from 50 percent to 100 percent across steady-states. All other
parameters are held fixed at their baseline values. This shock is unanticipated prior to time
15
t = 1. This amounts to reducing dij − 1 by 55 percent for each bilateral trade pair.
4.1 Welfare gains from trade
We compute the steady-state gains from trade using equation (2) and the dynamic gains from
trade using equation (1). We find that the steady-state gains from trade vary substantially
across countries, ranging from 18 percent for the U.S. to 92 percent for Belize. The median
gain (Greece) is 53 percent.
Dynamic gains for the median country (Greece) are 32 percent. The differences are large
across countries, ranging from 11 percent for the U.S., to 56 percent for Belize.9
The distribution of the dynamic gains from trade looks almost identical to the distribution
of the steady-state gains. However, the dynamic gains are smaller in each country. The
average ratio of dynamic gains to steady-state gains is 60.2 percent across countries, and
varies from a minimum of 60.1 percent to a maximum of 60.5 percent.10
The proportionality of roughly 60 percent is a result of (i) the initial change in con-
sumption and (ii) the rate at which consumption converges to the new steady-state. If
consumption jumped to its new steady-state level on impact then this ratio would be close
to 100 percent. If instead consumption declined significantly in the beginning, and then
converged to the new steady-state after many years, then the ratio could be closer to 0 per-
cent since there would be consumption losses in earlier periods while higher levels of future
consumption would be discounted.
The Euler equation reveals the forces that influence consumption dynamics. Trade lib-
eralization improves each country’s terms of trade making more resources available for both
consumption and investment. The allocation of output to consumption and investment
is determined optimally by the household. The relative price of investment falls, mean-
ing that the household can increase investment by a larger proportion than the increase
in output without giving up consumption. In addition, the marginal product of capital
(MPK),(
1 + rit+1
Pxit+1− δ)
, is higher than the steady-state MPK, 1β. While the MPK is high,
households take advantage by investing relatively more. Figure 2 shows the transition paths
for the relative price of investment and the MPK in the U.S. Resultantly, consumption falls
9The gains from trade are systematically smaller for large countries, rich countries, and countries withsmaller average export barriers. All of these findings are consistent with the existing literature (Waugh andRavikumar, 2016; Waugh, 2010).
10Desmet, Nagy, and Rossi-Hansberg (2015) consider, in a model of migration and trade, a counterfactualscenario that increases trade costs by 40% in the first period. They find that welfare decreases by around34%.
16
on impact and investment jumps as shown in Figure 3. As capital accumulates the MPK
returns to its original steady-state level and investment settles down to its new (higher)
steady-state level.11
Figure 2: Transition paths for prices in the U.S.
(a) Relative price of investment
0 20 40 60 80 100
Years after liberalization
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
Initial steady state
Transition
(b) Marginal product of capital
0 20 40 60 80 100
Years after liberalization
0.9
0.95
1
1.05
1.1
1.15
Initial steady state
Transition
4.2 The mechanism
Some remarks are in order here regarding the importance of two features that distinguish
our work from the literature: the endogenous relative price of capital and the endogenous
investment rate. In our model, the share of income that the household allocates towards
investment expenditures is determined endogenously. The nominal investment rate, PxitXitLiPcityit
is not constant along the transition path. Combined with a decline the relative price of
investment, the real investment rate, XitLiyit
, increases substantially in response to trade liber-
alization. Indeed, the real investment rate is permanently higher.
Alternative models To quantify the importance of the endogenous investment rate
and endogenous relative price of investment, we solve versions of the model where we explic-
11Two housekeeping remarks are in order here. First, in the figures we index each series to 1 in the initialsteady-state. Second, the transition paths for every country exhibit similar characteristics to the U.S., butdiffer in their magnitudes: Belize is at one extreme and the U.S. is at the other extreme.
17
Figure 3: Transitions paths for final demand in the U.S.
(a) Consumption
0 20 40 60 80 100
Years after liberalization
1
1.2
1.4
1.6
1.8
Initial steady state
Transition
(b) Investment
0 20 40 60 80 100
Years after liberalization
1
2
3
4
5
Initial steady state
Transition
itly impose that Px/Pc = 1 and/or that the nominal investment rate is exogenous. To do
this we change only a couple of equations.
First, to impose an exogenous nominal investment rate, we eliminate the Euler equation
from the baseline model and impose PxitXit = ρ(witLit + ritKit), with ρ = αδ1/β−(1−δ) =
0.1948. That is, the household allocates an exogenous share, ρ, of its income to investment
expenditures. The value of ρ corresponds to the nominal investment rate that arises in the
fully endogenous model in the steady-state (which is constant across countries and across
steady-states).
We implement a similar trade liberalization in which barriers are uniformly reduced by
55 percent in every country. We report the numbers for Greece only, since it is the country
that has the median gains from trade. All of the conclusions that we draw from Greece hold
in every other country.
We find that the endogenous investment rate affects the speed of capital accumulation.
For instance, Figure 4 shows that capital converges faster to the new steady-state in the
model with an endogenous investment rate. Indeed Table 2 reports the half-life for capital
accumulation: it is about twice as large in the model with an exogenous investment rate.
Second, in addition to an exogenous nominal investment rate, we fix the relative price of
investment to one. To do this we restrict the technologies for consumption and investment
goods to be the same. That is, we set Axi = Aci and νx = νc. In the calibration we choose
18
Ax and Ac to match the price of GDP relative to intermediates, and choose νx = νc = 0.88
to satisfy the national account equation (3), with all other parameters recalibrated to match
the same targets as in the benchmark calibration.
Again, we implement a similar trade liberalization in which barriers are uniformly re-
duced by 55 percent in every country and report the results for Greece. We find that an
endogenous relative price governs the gap in capital between steady-states. For instance,
Figure 4 shows that, with a fixed relative price of investment, the capital stock converges
to a lower steady-state level. Indeed, having an endogenous relative price allows for higher
steady-state capital-output ratio. With the relative price of investment fixed, the real invest-
ment rate, XiyiLi
, cannot adjust across steady-states since ρ = PxiXiPciyiLi
= αδ1/β−(1−δ) is constant.
On the other hand, with an endogenous relative price, the real investment rate converges to
a higher steady-state level since the opportunity cost of investing is lower, i.e., the amount
of consumption goods the household gives up to acquire additional investment is lower.
Figure 4: Transitional dynamics for capital across alternative models
. We consider Greecesince it is the country with the median gains from trade.
In sum, an endogenous investment rate allows the economy to transition to the steady-
state faster, while an endogenous relative price allows the economy to attain higher steady-
state capital stocks. These features have implications for the path of consumption along the
19
transition, and hence, for the ratio of dynamic-to-steady-state gains from trade.
The ratio of dynamic-to-steady-state gains is a function of (i) the initial change in con-
sumption and (ii) the rate of consumption growth, which depends on the half-life for capital.
Table 2 shows that the half-life for capital does not depend critically on whether the relative
price is fixed or not. For instance, in the model with an exogenous investment rate and
fixed relative price the half-life is 18.2 years. In the model with an exogenous investment
rate and endogenous relative price the half-life is 19.5 years. However, the initial change in
consumption depends on whether the relative price is fixed or not. In particular, when the
relative price is fixed, consumption increases by 13.1 percent on impact, whereas it increases
by only 9.9 percent when the relative price is not fixed. As a result, the ratio of dynamic-
to-steady-state gains is higher in the model with fixed relative price of investment.12
Conversely, in the model with endogenous relative price of investment, regardless of
whether the investment rate is exogenous of endogenous, the ratio of dynamic-to-steady-
state gains are similar. When the nominal investment rate is exogenous, the half-life is
twice as large as in the model with endogenous investment rate, but the initial increase
in consumption is higher. However, when the nominal investment rate is endogenous, the
half-life is shorter, but consumption drops on impact. In other words, with endogenous
investment rate, consumption is lower in the beginning of the transition, but converges to
the new steady-state faster.
Consequently, the welfare gains from trade that accounts for the whole transition path
of the economy in a model where both the investment rate and the relative price of capital
are endogenous, are different from models that take one or both of them as exogenous.
Table 2: Outcomes in Greece from global 55% reduction in barriers
Fixed inv.Baseline Fixed inv. rate + rel. price
Half life for capital 9.9 yrs 19.5 yrs 18.2 yrsInitial change consumption -5.1 % 9.9 % 13.2 %Dynamic-to-SS gains 60.4 % 59.6 82.4 %
Notes: We consider Greece since it is the country with the median gains
from trade.
12Both the dynamic and steady-state gains from trade liberalization are lower in the model with the fixedrelative price, but the ratio of the two is higher than in the model with endogenous relative price.
20
4.3 Putting the dynamic gains in perspective
In this section we perform two exercises. First, we compare dynamic welfare gains from trade
in a model with capital accumulation to the static gains that would be obtained in a model
with no capital accumulation. Then we relate our methodology to a sufficient-statistics
approach in which gains from trade are explained by changes in the home trade share.
Static versus dynamic welfare gains Here we compare our dynamic gains from
trade to those that would be obtained in a model with no capital accumulation (i.e., Waugh,
2010). In a static model, the welfare gains from trade are driven entirely by changes in TFP.
In Appendix D we show that the steady-state income per capita (which, recall is proportional
to consumption per capita) can be expressed as
yi ∝ Aci
(Tmiπii
) 1−νcθνm
︸ ︷︷ ︸TFP contribution
Aα
1−αxi
(Tmiπii
) α(1−νx)(1−α)θνm
︸ ︷︷ ︸Capital contribution
(8)
Equation (8) allows us to tractably decompose the relative importance of changes in TFP
and changes in capital in accounting for the gains. It implies that the log-change in income
that corresponds to a log-change in the home trade share is:
∂ ln(yi)
∂ ln(πii)= −
1− νcθνm︸ ︷︷ ︸
through TFP
+α(1− νx)
(1− α)θνm︸ ︷︷ ︸through capital
(9)
Based on our calibration, the first term equals 0.08 while the second term equals 0.30. That
is, given a change in trade barriers, 79 percent of the resulting change in income per capita
across steady-states can be attributed to change in capital, and the remaining 21 percent to
change in TFP.13
After a trade liberalization, TFP jumps immediately to its new steady-state level. Be-
cause the stock of capital does not change on impact, the initial change in TFP is not
affected by whether or not there is capital accumulation. Therefore, the initial change in
13This number is constant across countries in our model since the elasticities (θ, α, νc, νm, νx) are allconstant across countries. This does not imply that income per capita changes by equal proportions acrosscountries, only that the relative contributions from TFP and capital are the same.
21
TFP corresponds to the static welfare gains in a model without capital, or one in which
capital is exogenous. As a result, the static gains are 21 percent of the steady-state gains.
We also know from our counterfactual exercise that dynamic gains are around 60 percent of
the steady-state gains in a model with capital accumulation. Therefore, the dynamic gains
are almost three times larger than static gains obtained by ignoring changes in capital.
A sufficient-statistics approach We compare, period by period, welfare gains from
trade using the same formula as in Arkolakis, Costinot, and Rodrıguez-Clare (2012) in
our model with capital (augmented ACR), to those resulting from comparing consumption
growth period by period. The first measure is a“sufficient statistics” calculation in that it
depends only on changes in the home trade share and elasticity parameters (to see why,
recall equation (8)).
In the steady-state, all the change in income per capita resulting from changes in trade
barriers are manifested in the home trade share as in ACR, augmented by the fact that
capital is endogenous and it depends on trade barriers as in Anderson, Larch, and Yotov
(2015) and Mutreja, Ravikumar, and Sposi (2014). The sufficient-statistics calculation is
equivalent to comparing welfare in a series of static exercises.
The second measure captures the effect of capital accumulation on welfare gains from
trade, and hence accounts for all of the transitional dynamics following a trade liberalization,
which is not reflected in the transition path of the home trade share. Figure 5 plots both
measures.
Feeding in the transition path for the home trade share, the augmented ACR formula
would imply that all the gains from trade occur in the first period. The reason is that welfare
gains occur through a decrease in the home trade share, which jumps upon impact and it
reaches its new steady-state immediately. This is consistent with models that measure welfare
gains from trade in a static context. If instead we take into account the transitional dynamics
and compute consumption growth period by period, we observe that consumption drops upon
impact. But, after the initial period, consumption growth is positive, and converges toward
zero as the economy reaches the new steady-state.
As a final note, the sufficient-statistics formula is typically applied to assess the welfare
costs of moving to autarky, since the home trade share in autarky is 1, and the current home
trade share is observed in the data. In moving to free trade, even in a static model, one
needs to solve for the home trade shares that arise under free trade. In our model there is no
sufficient-statistic to compute the home trade share under free trade, even in steady-state,
22
Figure 5: Comparison of welfare measures along the transition
(a) Augmented ACR
0 20 40 60 80 100
Years after liberalization
1
1.1
1.2
1.3
1.4
1.5
(b) Consumption growth
0 20 40 60 80 100
Years after liberalization
0.96
0.98
1
1.02
1.04
1.06
since it depends on the world distribution of capital.14
This exercise shows that sufficient-statistics approaches can yield a very different picture
for the transitional dynamics of the welfare gains from trade, particularly when distinguishing
between the short-run gains and the long-run gains.
5 Extended model
Our algorithm for solving the baseline model encompasses multi-country models with CRRA
utility, linear depreciation of capital, and balanced trade. The main limitation is that house-
holds cannot borrow against the future because physical investment is the only means of
intertemporal substitution, and investment must be non-negative. This restricts us to study
transition paths in which investment is always positive. To get around the non-negativity
constraint we introduce adjustment costs to capital to ensure that the household will opti-
mally choose positive investment every period. To allow the household to borrow against
the future we introduce one-period bonds that can be used to finance trade imbalances. The
Appendix describes the algorithm to compute the equilibrium transitional dynamics in the
extended model.
14In a one-sector model with no capital, Waugh and Ravikumar (2016) derive sufficient statistics to measurethe gains from moving to free trade. It involves the elasticities, the current home trade share, and currentlevel of output.
23
5.1 Adjustment costs and endogenous trade imbalances
Adjustment costs are added to the technology for physical capital accumulation as follows
Kit+1 = (1− δ)Kit + χXµitK
1−µit
where χ denotes an adjustment cost, and µ denotes the rate at which investment goods
are converted into future capital stock. We work with the inverse capital accumulation
technology for convenience, which is given by
Xit = Φ(Kit+1, Kit) =
(1
χ
) 1µ
(Kit+1 − (1− δ)Kit)1µ K
µ−1µ
it
The Euler equation for investment in physical capital becomes
Cit+1 = βσ
( rit+1
Pixt+1− Φ2(Kit+2, Kit+1)
Φ1(Kit+1, Kit)
)σ (Pxit+1/Pcit+1
Pxit/Pcit
)σCit
where Φ1(·, ·) and Φ2(·, ·) denote the first derivatives of the adjustment-cost function with
respect to the first and second arguments, respectively:
Φ1(Kit+1, Kit) =
(1
χ
) 1µ(
1
µ
)(Kit+1
Kit
− (1− δ)) 1−µ
µ
Φ2(Kit+1, Kit) =
(1
χ
) 1µ(
1
µ
)(Kit+1
Kit
− (1− δ)) 1−µ
µ(
(µ− 1)Kit+1
Kit
− µ(1− δ))
The second extension we consider is adding endogenous trade imbalances, in addition to
adjustment costs to capital accumulation. We allow each country to freely trade one-period
bonds that yield a risk-free world interest rate qt. More specifically, the household’s budget
constraint becomes
PcitCit + PxitXit +Bit = witLit + ritKit + qtAit
where Bit denotes net purchases of one-period bonds (the current account balance) and Aitdenotes the net-foreign asset position. Net-foreign assets accumulate according to
Ait+1 = Ait +Bit
24
This adds one more Euler equation to the household’s optimization problem, namely
Cit+1
Cit= βσ
(1 + qt+1
Pcit+1/Pcit
)σTo close the model, instead of imposing balanced trade period-by-period, we require that
the current account equals net exports plus net-foreign income on assets:15
Bit = Pmit (Yit −Mit) + qtAit
5.2 Quantitative exercise
We calibrate the adjustment cost parameter, χ = δ1−µ so that there is no cost to mainiatin
the level of capital stock in steady-state, that is, Xi = δKi. We set the elasticity parameter
µ = 0.55 as in Eaton, Kortum, Neiman, and Romalis (2016).
To put the model to work, we assume the world is in steady-state in 2011 with balanced
trade and a balanced current account. In addition, we assume that the initial net-foreign
asset position is zero in every country: Ai1 = 0. We use the steady-state level of capital
stock that arises from this calibration as the initial level for capital. These assumptions are
made so that we can explore the dynamics along the transition path free from any differences
in initial conditions and make direct comparisons to the results in a model with balanced
trade in every period.
To implement our algorithm, we impose a value for the terminal net-foreign asset positions
(NFAPs), AiT+1 for all i; without loss of generality, we set these to zero. We let the model
run for 150 periods and discard the last 65. By period 85, the model has converged to a
steady-state that is independent from the terminal NFAPs. This is an application of the
Turnpike Theorem whereby, regardless of the terminal condition of the AiT+1, if T is large
enough, then there is a time t? at which the model is sufficiently close to the steady-state,
i.e., on the Turnpike. This is approach is also used by Kehoe, Ruhl, and Steinberg (2016).16
15With abuse of notation this implies, equivalently, that Y = C + I +NX.16In models with endogenous current accounts, it is known that the transition path and steady-state are
determined jointly. The new steady-state is one in which current accounts are balanced in all countriesbut have potentially permanent trade imbalances. The steady-state depends on the transition path sincea country may accumulate bonds early on, financed by a trade surplus, but later on collect income off ofthe assets and use it to finance a trade deficit. As such, its steady-state trade imbalance depends on whathappened along the transition. So net-foreign asset positions need not be zero in steady-state, but they doneed to be constant. After 85 periods, the net-foreign asset position begins to regress to its terminal position,i.e., the transition path exits the Turnpike.
25
Cross-sectional analysis We find that, after a trade liberalization, capital flows from
large and more developed countries to small and less-developed countries. The reason is
that when there is financial autarky (i.e., trade is balanced), the large countries have a
lower real rate of return than the small countries. That is, under balanced trade, there
are persistent differences in the real rate of return to capital (RRR) across countries along
the transition path, ranging from 4 percent to 7 percent. With trade imbalances, capital
flows from countries that have a lower marginal product of capital to countries that have
a higher marginal product of capital. This implies that, on impact, small countries run
current account deficits while large countries run current account surpluses, see Figure 6a.
As a result, RRRs are equalized across countries to 1/β = 4.17 percent (see Figure 6b).
Countries that initially run a deficit have a lower RRR relative to the model with balanced
trade; the opposite is true for countries in surplus.
Figure 6: Current account imbalances and real rates of return across countries in period 1
(a) Current account over GDP
0.1 1 10 100 1000 10,000 100,000
Real income, millions 2005 U.S. dollars
-0.1
-0.05
0
0.05
ARM
AUS
AUTBDI
BEN
BGD
BGRBHS
BLR
BLZ
BRA
BRB
BTN
CAF
CAN
CHECHL
CIVCMR
COL
CPVCRI
CYP
CZE
DEU
DNK
DOM
ECUEGY
ESP
ETH
FIN
FJI
FRAGBR
GEO
GRC
GTM
HND
HUNIDN
IND
IRL
IRN
ISL
ISR
ITA
JAM JOR
JPN
KGZKHM
KOR
LKALSO
MARMDA
MDG
MDV
MEX
MKDMOZMUS
MWI
NPLNZL
PAK
PERPHLPOL
PRT
PRY
ROU
RUSRWA
SEN
STP
SWETHA
TUN
TUR
TZA
UGA
UKR
URY
USA
VCT
VEN
VNM
YEM
ZAFBAL
CHM
(b) Real rate of return
-0.08 -0.06 -0.04 -0.02 0 0.02
Current account share in GDP in period 1
1.04
1.05
1.06
1.07
Model with endogenous current acoount
ARM
AUS
AUTBDI
BEN
BGD
BGR
BHS
BLR
BLZ
BRA
BRB
BTN
CAF
CAN
CHECHL
CIVCMR
COL
CPVCRICYP
CZE
DEU
DNK
DOM
ECU
EGY
ESP
ETH
FIN
FJI
FRAGBR
GEO
GRC
GTM
HND
HUN
IDN
IND
IRL
IRN
ISL
ISR
ITA
JAMJOR
JPN
KGZ
KHM
KOR
LKALSO
MARMDA
MDG
MDV
MEX
MKD
MOZMUS
MWI
NPL
NZL
PAK
PERPHL
POL
PRT
PRY
ROU
RUSRWA
SEN
STP
SWE
THA
TUN
TUR
TZA
UGA
UKR
URY
USA
VCT
VEN
VNM
YEM
ZAF
BAL
CHM
As the world transitions towards the new steady-state, each country’s current account
converges towards zero. Along the transition, some countries accumulate positive NFAPs
by running current account surpluses, while other accumulate negative NFAPs (liabilities)
by running current account deficits. To illustrate this point, consider the case of the U.S
and Belize. Following the trade liberalization, Figure 7a shows that the U.S. runs a current
account surplus. On impact, the current account surplus exactly equals the U.S. trade
surplus. In the ensuing periods, the trade surplus shrinks faster than the current account
26
surplus, since the U.S. earns positive net-foreign income off of its existing NFAP. In a matter
of 14 periods, the U.S. trade is balanced, meaning that net purchases of bonds (its current
account balance) exactly offsets its net-foreign income. After period 14, the U.S. continues
to run a current account surplus, although its net exports turn negative. In the new steady-
state, its current account is balanced and the U.S. runs a permanent trade deficit that is
1.5 percent of GDP as shown in Figure 7b. The U.S. trade deficit is financed by net-foreign
income that accrues off of its permanent and positive NFAP.
The current account dynamics in Belize are the mirror image of those in the U.S. whereas
Belize converges to a steady-state with permanent net liabilities offset by a permanent a trade
surplus.
Figure 7: Transition for current account and net exports in the U.S. and in Belize
(a) Current account over GDP
0 20 40 60 80
Years after liberalization
-0.05
0
0.05
BLZ
USA
(b) Net exports over GDP
0 20 40 60 80
Years after liberalization
-0.05
0
0.05
BLZ
USA
These results have implications for welfare since they impact the transition path for con-
sumption and capital accumulation. With trade imbalances, countries running a trade deficit
consume and invest more than they would if trade were balanced, and the opposite is true
for countries running a surplus. As a result, countries like Belize are able to accumulate
capital at a faster rate relative to the model with balanced trade, whereas the U.S. accumu-
lates capital at a relatively slower rate. Indeed, Figure 8 shows that the half-life for capital
accumulation is shorter for countries that initially run deficits compared to countries that
initially run surpluses.
The rate of capital accumulation and the dynamics of trade imbalances govern the tran-
27
Figure 8: Half-life for capital accumulation
-0.08 -0.06 -0.04 -0.02 0 0.02
Current account share in GDP in period 1
22
23
24
25
26
27
ARM
AUS
AUTBDI
BEN
BGDBGR
BHS
BLR
BLZ
BRA
BRBBTN
CAF
CAN
CHE
CHLCIV
CMRCOL
CPV
CRI
CYP
CZE
DEU
DNK
DOM
ECU
EGY
ESP
ETH
FIN
FJI
FRA
GBR
GEO
GRCGTM
HND
HUN
IDN
IND
IRL
IRN
ISLISR
ITA
JAM
JOR
JPN
KGZKHM
KOR
LKALSO
MARMDA MDGMDV
MEX
MKD
MOZMUSMWI
NPLNZL
PAK
PERPHL
POL
PRT
PRY
ROU
RUSRWA
SENSTP
SWE
THATUN
TUR
TZA
UGA
UKR
URY
USA
VCT
VEN
VNMYEM
ZAFBAL
CHM
sition for consumption. We find that countries that run trade deficits on impact have higher
initial consumption than they would in a model with balanced trade, while countries that run
trade surpluses have lower initial consumption, see Figure 9a. The trade-off is that, countries
that initially run deficits, like Belize, will eventually run trade surpluses. As a result, the
steady-state level of consumption for Belize is lower than in a model with balanced trade.
Conversely, countries that initially run surpluses, such as the U.S., will benefit by having
higher levels of steady-state consumption than in a model with balanced trade. On net, wel-
fare gains are slightly higher in every country, relative to the gains calculated in the model
with balanced trade. However, countries that initially run a deficit gain proportionately
more than countries that initially run a surplus.
Implications for sufficient-statistics approach Recall from section 4.3 that the
baseline model with balanced trade admits a sufficient-statistics formula to compute the
welfare gains from trade across steady-states, but not in the transition. In the model with
endogenous trade imbalances, the sufficient-statistics formula no longer applies, even between
steady-states. Specifically, equation (9) implies that the elasticity of the change in income
with respect to the change in the home trade share is constant across countries in the baseline
28
Figure 9: Percent change in consumption relative to model with balanced trade
(a) Change in period-1 consumption
-0.08 -0.06 -0.04 -0.02 0 0.02
Current account share in GDP in period 1
0
2
4
6
ARM
AUS
AUTBDI
BEN
BGD
BGRBHS
BLR
BLZ
BRA
BRB
BTN
CAF
CAN
CHECHL
CIVCMR
COL
CPV
CRI
CYP
CZE
DEU
DNK
DOM
ECU
EGY
ESP
ETH
FIN
FJI
FRA
GBR
GEO
GRC
GTM
HND
HUN
IDN
IND
IRL
IRN
ISL
ISR
ITA
JAMJOR
JPN
KGZ
KHM
KOR
LKA
LSO
MARMDA
MDG
MDV
MEX
MKD
MOZMUS
MWI
NPL
NZL
PAK
PERPHL
POL
PRT
PRY
ROU
RUSRWA
SEN
STP
SWE
THA
TUN
TUR
TZA
UGA
UKR
URY
USA
VCT
VEN
VNM
YEM
ZAF
BAL
CHM
(b) Change in steady-state consumption
-0.08 -0.06 -0.04 -0.02 0 0.02
Current account share in GDP in period 1
-8
-6
-4
-2
0
2
ARM
AUS
AUTBDI
BEN
BGD
BGRBHS
BLR
BLZ
BRA
BRB
BTN
CAF
CAN
CHECHL
CIV
CMR
COL
CPV
CRICYP
CZE
DEU
DNK
DOM
ECU
EGY
ESP
ETH
FIN
FJI
FRA
GBR
GEO
GRC
GTM
HND
HUN
IDN
IND
IRL
IRN
ISL
ISR
ITA
JAMJOR
JPN
KGZ
KHM
KOR
LKALSO
MARMDA
MDG
MDV
MEX
MKD
MOZMUS
MWI
NPL
NZL
PAK
PERPHL
POL
PRT
PRY
ROU
RUSRWA
SEN
STP
SWE
THA
TUN
TUR
TZA
UGA
UKR
URY
USA
VCT
VEN
VNM
YEM
ZAF
BAL
CHM
model. Nonetheless, Figure 10 shows that this elasticity systematically varies across countries
with the level of the trade imbalance in the extended model.
The real income in a country is affected by its steady-state trade imbalance, a phe-
nomenon that is not fully summarized by the home trade share alone. Countries that run
steady-state trade deficits, like the U.S., have lower (in absolute value) elasticities than those
that run steady-state trade surpluses, like Belize. This implies that, given the same drop
in the home trade share, the steady-state change in income per capita will be higher in
Belize than in the U.S. Since the home trade share decreases more in Belize than in the
U.S., the steady-state gains are proportionately larger in Belize than in the U.S., relative to
a model with trade balance. Since the steady-state outcomes depend on the entire transition
in the model with trade imbalances, it is important to take into account to entire transi-
tion path when evaluating the gains from trade, even if one is interested in the gains across
steady-states only.
29
Figure 10: Elasticity of change in income with respect to change in home trade share
-0.01 0 0.01 0.02 0.03 0.04 0.05
Net exports over GDP in new steady state
-0.38
-0.37
-0.36
-0.35
-0.34
ARM
AUS
AUTBDI
BEN
BGDBGR
BHS
BLR
BLZ
BRA
BRBBTN
CAF
CAN
CHECHLCIVCMR
COL
CPVCRI CYP
CZE
DEU
DNKDOM
ECUEGY
ESP
ETHFIN
FJI
FRAGBR
GEO
GRCGTM
HND
HUNIDN
IND
IRL
IRN
ISLISR
ITA
JAMJOR
JPN
KGZKHM
KOR
LKALSOMARMDAMDG
MDV
MEX
MKDMOZMUSMWI
NPLNZL
PAK
PERPHLPOL
PRT
PRYROU
RUSRWA
SENSTP
SWETHA
TUN
TUR
TZA
UGA
UKRURY
USA
VCT
VENVNM
YEM
ZAF BAL
CHM
Model with balanced trade
Model with unbalanced trade
6 Conclusion
We build a multi-country trade model with capital accumulation to study the welfare gains
from trade. The model features endogenous investment rate and endogenous relative price
of investment. We then solve for the exact transitional dynamics of a trade liberalization in
levels. Our counterfactual suggests that dynamic gains are 60 percent of the gains across
steady-states, and three times larger than those implied by a static model with no capital
accumulation. Furthermore, endogenous relative price of investment implies higher steady-
state capital stocks whereas endogenous investment rate implies faster convergence towards
the steady-state.
Our paper adds to the literature on measuring welfare gains from trade, and more gen-
erally, to the recent literature addressing dynamics in multi-country models of trade. Many
of the static models are based on “sufficient statistics”. We find large difference between
changes in welfare in a model with endogenous capital accumulation and those measured by
“sufficient statistics” along the transition, pointing to the importance on modeling dynamics
explicitly when measuring the benefits of openness.
Finally, we extend our model by adding adjustment costs to capital accumulation and
30
endogenous trade imbalances to allow for intertemporal borrowing. This version of the model
implies that after a trade liberalization, small countries run trade deficits in the short run
and accumulate capital faster than large countries. The small countries, however, run trade
surpluses in the long run, but have proportionately larger dynamic welfare gains relative to
a model with balanced trade.
Our paper emphasizes the importance of analyzing the entire transitional dynamics of
the economies after a trade liberalization to measure welfare gains from trade. These results
have implications for the effect of trade policy on economic growth, which we leave for future
work.
References
Affendy, Arip M, Lau Sim Yee, and Madono Satoru. 2010. “Commodity-industry Classifi-
cation Proxy: A Correspondence Table Between SITC Revision 2 and ISIC Revision 3.”
MPRA Paper 27626, University Library of Munich, Germany.
Alessandria, George, Horag Choi, and Kim Ruhl. 2015. “Trade Adjustment Dynamics and
the Welfare Gains from Trade.” Working paper.
Alvarez, Fernando and Robert E. Lucas. 2007a. “Capital Accumulation and International
Trade.” Working paper, University of Chicago.
———. 2007b. “General Equilibrium Analysis of the Eaton-Kortum Model of International
Trade.” Journal of Monetary Economics 54 (6):1726–1768.
Anderson, James, Mario Larch, and Yoto Yotov. 2015. “Growth and Trade with Frictions: A
Structural Estimation Framework.” Working Paper 21377, National Bureau of Economic
Research.
Arkolakis, Costas, Arnaud Costinot, and Andres Rodrıguez-Clare. 2012. “New Trade Models,
Same Old Gains?” American Economic Review 102 (1):94–130.
Bernard, Andrew B, Jonathan Eaton, J Bradford Jensen, and Samuel Kortum. 2003. “Plants
and Productivity in International Trade.” American Economic Review 93 (4):1268–1290.
Brooks, Wyatt J. and Pau S. Pujolas. 2016. “Capital Acumulation and the Welfare Gains
from Trade.” Working Paper Series 2016-03, McMaster University.
31
Caliendo, Lorenzo, Maximiliano Dvorkin, and Fernando Parro. 2015. “The Impact of Trade
on Labor Market Dynamics.” Working Paper 21149, National Bureau of Economic Re-
search.
Desmet, Klaus, David Krisztian Nagy, and Esteban Rossi-Hansberg. 2015. “The geography
of development: Evaluating migration restrictions and coastal flooding.” .
Eaton, Jonathan and Samuel Kortum. 2002. “Technology, Geography, and Trade.” Econo-
metrica 70 (5):1741–1779.
Eaton, Jonathan, Samuel Kortum, Brent Neiman, and John Romalis. 2016. “Trade and the
Global Recession.” American Economic Review 106 (11):3401–3438.
Feenstra, Robert C., Robert Inklaar, and Marcel Timmer. 2015. “The Next Generation of
the Penn World Table.” American Economic Review 105 (10):3150–82.
The left-hand side denotes the gross output of intermediates in country i and the right-hand
side denotes total expenditures on intermediates.
B Solution algorithm for the baseline model
In this section of the Appendix we describe the algorithm for computing 1) the steady-state
and 2) the transition path. Before going further into the algorithms, we introduce some
notation. We denote the steady-state objects using the ? as a superscript, i.e., K?i is the
steady-state stock of capital in country i. We denote the cross-country vector of capital at
a point in time using vector notation; ~Kt = {Kit}Ii=1 is the vector of capital stocks across
countries at time t.
B.1 Computing the steady-state equilibrium in the baseline model
The steady-state equilibrium consists of 23 objects: ~w?, ~r?, ~P ?c , ~P ?
m, ~P ?x , ~C?, ~X?, ~K?, ~Q?, ~Y ?
c ,~Y ?m, ~Y ?
x , ~K?c , ~K?
m, ~K?x, ~L?c , ~L
?m, ~L?x, ~M
?c , ~M?
m, ~M?x , ~~π?. Table B.1 provides a list of equilibrium
conditions that these objects must satisfy.
We use the technique from Mutreja, Ravikumar, and Sposi (2014), which builds on
Alvarez and Lucas (2007b), to solve for the steady-state. The idea is to guess at a vector
of wages, then recover all remaining prices and quantities using optimality conditions and
market clearing conditions, excluding the trade balance condition. We then use departures
from the the trade balance condition in each country to update our wage vector and iterate
until we find a wage vector that satisfies the trade balance condition. The following steps
outline our procedure in more detail.
1. We guess a vector of wages ~w ∈ ∆ = {w ∈ RI+ :∑I
i=1wiLi1−α = 1}; that is, with world
GDP as the numeraire.
2. We compute prices ~Pc, ~Px, ~Pm, and ~r simultaneously using conditions 16, 17, 18, and
23 in Table B.1. To complete this step, we compute the bilateral trade shares ~~π using
condition 19.
3. We compute the aggregate capital stock as Ki = α1−α
wiLiri
, for all i, which derives easily
from optimality conditions 1 & 4, 2 & 5, and 3 & 6, coupled with market clearing
conditions for capital and labor 10 &11 in Table B.1.
36
Table B.1: Equilibrium conditions in steady-state
1 r?iK?ci = ανcP
?ciY
?ci ∀(i)
2 r?iK?mi = ανmP
?miY
?mi ∀(i)
3 r?iK?xi = ανxP
?xiY
?xi ∀(i)
4 w?iL?ci = (1− α)νcP
?ciY
?ci ∀(i)
5 w?iL?mi = (1− α)νmP
?miY
?mi ∀(i)
6 w?iL?xi = (1− α)νxP
?xiY
?xi ∀(i)
7 P ?miM
?ci = (1− νc)P ?
ciY?ci ∀(i)
8 P ?miM
?mi = (1− νm)P ?
miY?mi ∀(i)
9 P ?miM
?xi = (1− νx)P ?
xiY?xi ∀(i)
10 K?ci +K?
mi +K?xi = K?
i ∀(i)11 L?ci + L?mi + L?xi = Li ∀(i)12 M?
ci +M?mi +M?
xi = Q?i ∀(i)
13 C?i = Y ?
ci ∀(i)14
∑Ij=1 P
?mj
(M?
cj +M?mj +M?
xj
)πji = P ?
miY?mi ∀(i)
15 X?i = Y ?
xi ∀(i)
16 P ?ci =
(1Aci
)(r?iανc
)ανc ( w?i(1−α)νc
)(1−α)νc ( P ?mi1−νc
)1−νc∀(i)
17 P ?mi = γ
[∑Ij=1(u?mjdij)
−θTmj
]− 1θ ∀(i)
18 P ?xi =
(1Axi
)(r?iανx
)ανx ( w?i(1−α)νx
)(1−α)νx ( P ?mi1−νx
)1−νx∀(i)
19 π?ij =(u?mjdij)
−θTmj∑Ij=1(u?mjdij)
−θTmj∀(i, j)
20 P ?miY
?mi = P ?
miQ?i ∀(i)
21 P ?ciC
?i + P ?
xiX?i = r?iK
?i + w?iL
?i ∀(i)
22 X?i = δK?
i ∀(i)23 r?i =
(1β− (1− δ)
)P ?xi ∀(i)
Note: u?mj =(
r?jανm
)ανm ( w?j(1−α)νm
)(1−α)νm ( P ?mj1−νm
)1−νm.
37
4. We use condition 22 to solve for steady-state investment ~X. Then we use condition 21
to solve for steady-state consumption ~C.
5. We combine conditions 4 & 13 to solve for ~Lc, combine conditions 5 & 14 to solve for~Lx, and use condition 11 to solve for ~Lm. Next we combine conditions 1 & 4 to solve for~Kc, combine conditions 2 & 5 to solve for ~KM , and combine conditions 3 & 6 to solve
for ~Kx. Similarly, we combine conditions 4 & 7 to solve for ~Mc, combine conditions 5
& 8 to solve for ~Mm, and combine conditions 6 & 9 to solve for ~Mx.
6. We compute ~Yc using condition 13, compute ~Ym using condition 14, and compute ~Yx
using condition 15.
7. We compute an excess demand equation as in Alvarez and Lucas (2007b) defined as
Zi(~w) =PmiYmi − PmiQi
wi
(the trade deficit relative to the wage). Condition 20 requires that Zi(~w) = 0 for all
i. If the excess demand is sufficiently close to zero then we have an equilibrium. If
not, we update our guess at the equilibrium wage vector using the information in the
excess demand as follows.
Λi(~w) = wi
(1 + ψ
Zi(~w)
Li
)is be the updated guess to the wage vector, where ψ is chosen to be sufficiently small
so that Λ > 0. Note that∑I
i=1Λi(~w)Li
1−α =∑I
i=1wiLi1−α + ψ
∑Ii=1 wiZi(~w). As in Alvarez
and Lucas (2007b), it is easy to show that∑I
i=1 wiZi(~w) = 0 which implies that∑Ii=1
Λi(~w)Li1−α = 1, and hence, Λ : ∆ → ∆. We return to step 2 with our updated
wage vector and repeat the steps. We iterate through this procedure until the excess
demand is sufficiently close to zero. In our computations we find that our preferred
convergence metric:I
maxi=1{|Zi(~w)|}
converges roughly monotonically towards zero.
38
Table B.2: Equilibrium conditions along the transition
using variables pl c and pl i respectively. These correspond to Pc and Px in our model.
We construct the price of intermediate goods (manufactures) by combining disaggre-
gate price data from the World Bank’s 2011 International Comparison Program (ICP):
http://siteresources.worldbank.org/ICPEXT/Resources/ICP 2011.html. The data has sev-
eral categories that fall under what we classify as manufactures: “Food and nonalcoholic
beverages”, “Alcoholic beverages, tobacco, and narcotics”, “Clothing and foot wear”, and
“Machinery and equipment”. The ICP reports expenditure data for these categories in both
nominal U.S. dollars and real U.S. dollars. The conversion from nominal to real uses the PPP
price, that is: the PPP price equals the ratio of nominal expenditures to real expenditures.
As such, we compute the PPP for manufactures as a whole of manufactures for each country
as the sum of nominal expenditures across categories divided by the sum of real expenditures
across categories. For the RoW aggregate, we simply sum the expenditure across all of the
countries that are not part of the 40 individual countries.
There is one more step before we take these prices to the model. The data correspond
to expenditures, thus include additional margins such as distribution. In order to adjust for
this this, we first construct a price for distribution services. We assume that the price of
distribution services is proportional to the overall price of services in each country and use the
same method as above to compute the price across the following categories: “Housing, water,
electricity, gas, and other fuels”, “Health”, “Transport”, “Communication”, “Recreation and
culture”, “Education”, “Restaurants and hotels”, and “Construction”.
Now that we have the price of services in hand, we strip it away from the price of goods
computed above to arrive at a measure of the price of manufactures that better corresponds
to our model. In particular, let Pd denote the price of distribution services and let Pg denote
the price of goods that includes the distribution margin. We assume that Pg = Pψd P
1−ψm ,
where Pm is the price of manufactures. We set ψ = 0.45 which is a value commonly used in
the literature.
F Additional figures and tables
Table F.1: Gains from trade (%) following uniform reduction in barriers by 55%
Model 1 Model 2 Model 3 Model 4 Model 5Country ISO Dyn SS Dyn SS Dyn Dyn Dyn SSArmenia ARM 24.2 29.3 46.9 77.8 47.2 35.8 36.0 61.7Australia AUS 10.9 13.2 20.5 34.5 20.8 16.2 16.2 31.6Austria AUT 14.3 17.3 26.5 44.6 26.9 20.8 20.8 38.7Continued on next page. . .
Model 1 Model 2 Model 3 Model 4 Model 5Country ISO Dyn SS Dyn SS Dyn Dyn Dyn SSParaguay PRY 21.5 26.0 40.8 67.4 40.9 31.3 31.4 54.8Peru PER 14.9 18.0 27.6 46.2 27.9 21.5 21.6 39.8Philippines PHL 16.0 19.3 29.3 48.8 29.5 22.8 22.9 41.8Poland POL 14.0 17.0 26.0 43.8 26.4 20.4 20.4 38.1Portugal PRT 16.3 19.7 30.3 50.7 30.7 23.6 23.7 43.0Rep. of Korea KOR 13.3 16.1 23.9 40.0 24.2 18.8 18.8 35.6Rep. of Moldova MDA 20.9 25.3 40.0 66.8 40.5 30.8 31.0 54.0Romania ROU 17.4 21.1 32.4 54.2 32.8 25.1 25.3 45.4Russian Federation RUS 12.5 15.2 23.4 39.6 23.9 18.4 18.5 35.1Rwanda RWA 11.8 14.3 22.0 37.3 22.5 17.4 17.4 33.5Saint Vincent and the Grenadines VCT 28.1 34.2 56.0 92.7 56.4 42.3 42.7 71.9Sao Tome and Principe STP 20.8 25.3 39.4 65.8 39.9 30.3 30.5 53.4Senegal SEN 23.5 28.5 46.2 76.8 46.6 35.3 35.5 61.0South Africa ZAF 15.3 18.6 29.3 49.0 29.6 22.8 22.9 41.8Southeast Europe SEE 17.4 21.1 32.4 54.4 32.9 25.2 25.3 45.5Spain ESP 11.7 14.2 21.4 36.3 21.9 16.9 17.0 32.8Sri Lanka LKA 17.4 21.1 32.1 53.6 32.4 24.9 25.0 45.1Sweden SWE 14.8 17.9 27.7 46.5 28.1 21.6 21.7 40.0Switzerland CHE 15.4 18.7 28.4 47.5 28.7 22.1 22.2 40.8TFYR of Macedonia MKD 20.3 24.6 38.4 64.2 38.9 29.6 29.8 52.3Thailand THA 17.4 21.0 32.2 53.6 32.4 25.0 25.0 45.1Tunisia TUN 21.2 25.7 40.3 67.1 40.7 31.0 31.1 54.4Turkey TUR 14.4 17.5 26.3 44.4 26.8 20.6 20.7 38.5USA USA 5.7 6.9 10.6 18.3 11.0 8.5 8.5 20.0Uganda UGA 11.8 14.3 22.0 37.4 22.5 17.4 17.5 33.5Ukraine UKR 15.9 19.3 30.3 50.9 30.7 23.6 23.7 43.0United Kingdom GBR 12.7 15.4 23.4 39.5 23.8 18.4 18.5 35.1United Rep. of Tanzania TZA 22.5 27.3 43.7 72.7 44.1 33.5 33.7 58.1Uruguay URY 19.0 23.0 36.6 60.9 36.9 28.3 28.4 50.1Venezuela VEN 15.5 18.7 29.1 48.4 29.2 22.6 22.7 41.5Viet Nam VNM 19.7 23.9 36.3 60.2 36.5 28.0 28.1 49.7Yemen YEM 22.1 26.9 41.9 69.3 42.0 32.1 32.2 56.0
Note: “Dyn” refers to dynamic gains and “SS” refers to steady-stategains. Model 1 is the model with exogenous nominal investment rate,fixed relative price of investment. Model 2 adds the endogenousrelative price of investment to model 1. Model 3 (baseline) addsthe endogenous nominal investment rate to Model 2. Model 4 addsadjustment costs to capital accumulation to Model 3. Model 5 addsthe endogenous trade imbalances to Model 4. Steady-state gains areidentical in Models 2,3, and 4. The group “Southeast Europe” is anaggregate of Albania, Bosnia and Herzegovina, Croatia, Montenegro,and Serbia.